bryancall commented on issue #9422: URL: https://github.com/apache/trafficserver/issues/9422#issuecomment-4828017981
This is a configuration conflict rather than a defect in the per-remap parameter. In the posted setup slice.so is loaded twice: globally in plugin.config (--blockbytes=1m) and per-remap in remap.config (-b 2097152, that is 2MB). When slice is loaded globally it installs a hook that runs on every transaction using the global configuration (1MB) and slices the incoming request into 1MB blocks before the per-remap instance is consulted. The two are separate plugin instances with separate configurations, so the global 1MB wins, which matches the 1048576-byte blocks shown in squid.log. The per-remap -b / --blockbytes parameter does work on its own (both the -b 2M and --blockbytes=2M forms are documented in plugins/slice/README.md, and 2MB is within the accepted 256KB to 128MB range). To get 2MB blocks for this map, configure slice in only one place for that path: remove the global slice.so line from plugin.config and keep the per-remap @pparam=-b @pparam=2097152 (or @pparam=--blockbyte s=2M). I am closing this as working as designed. Please reopen with details if 2MB blocks are still not produced after removing the global slice entry. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
